Allows overriding settings via the global 'ssl' settings as before.
This order was probably accidental. That said, 'ssl' is a giant footgun
we will want to discourage use of.
These provide (a) a way to deal with random assortments of certs
and (b) avoid unnecessary error messages and warnings, according
to #1669 anyway, which this fixes.
The goal is to allow module:provides("foo-bar") with a mod_foo_bar_ prefix
being stripped. It will break any existing modules that use a prefix and have
hyphens instead of underscores. No such modules are known.
The existing events do not fire for unauthed sessions, for example (because
the type does not match). I deemed changing their behaviour too risky, and
the current behaviour may even be more desirable for some uses.
This means we now have roughly paired events:
- s2s-created -> s2s-destroyed (global only)
- s2sin-established -> s2sin-destroyed (global + host)
- s2sout-established -> s2sout-destroyed (global + host)
To prevent a situation where you for whatever reason use a full JID that
is currently online and the response ends up routed there instead of the
module:send_iq() handlers.
This is primarily something that happens with an internal query to
mod_mam, which calls origin.send() several times with results, leading
to the first such result being treated as the final response and
resolving the promise.
Now, these responses pass trough to the underlying origin.send(), where
they can be caught. Tricky but not impossible. For remote queries, it's
even trickier, you would likely need to bind a resource or similar.
Removes the need to enable DANE with two separate settings.
Previously you had to also set `ssl = { dane = true }` to activate DANE
support in LuaSec and OpenSSL.
Quick Fix\u{2122} to stop prevent certmanager from automatically adding
a client certificate for net.http.request, since this normally does not
require such.
Under some circumstances when hosts and modules are loaded in some
certain order, entries end up missing from the SNI map. This manifests
in e.g. `curl https://localhost:5281/` giving an error about
"unrecognized name".
The `service` argument is `nil` when invoked from the "host-activated"
event, leading it to iterating over every service. And then it would not
be fetching e.g. `http_host` from the config, which explains why https
would sometimes not work due to the missing name entry.
Because when `service` is included, this limits the iteration to
matching entries, while also returning the same value as the `name` loop
variable. Because `name == service when service != nil` we can use name
instead in the body of the loop.
This was a leftover from when we (or rather I) thought that the
old (now called "high-level") API would be removed. We deemed it
useful though, so let's remove that "legacy" language and make
the description more friendly.
lfs.dir() throws a hard error if there's a problem, e.g. no such
directory or permission issues. This also gets called early enough that
the main loop error protection hasn't been brought up yet, causing a
proper crash.
Otherwise the default "certs" would be relative to $PWD, which works
when testing from a source checkout, but not on installed systems where
it usually points to the data directory.
Also, the LuaFileSystem dir() iterator throws a hard error, which may
cause a crash or other problems.
The metric subsystem of Prosody has had some shortcomings from
the perspective of the current state-of-the-art in metric
observability.
The OpenMetrics standard [0] is a formalization of the data
model (and serialization format) of the well-known and
widely-used Prometheus [1] software stack.
The previous stats subsystem of Prosody did not map well to that
format (see e.g. [2] and [3]); the key reason is that it was
trying to do too much math on its own ([2]) while lacking
first-class support for "families" of metrics ([3]) and
structured metric metadata (despite the `extra` argument to
metrics, there was no standard way of representing common things
like "tags" or "labels").
Even though OpenMetrics has grown from the Prometheus world of
monitoring, it maps well to other popular monitoring stacks
such as:
- InfluxDB (labels can be mapped to tags and fields as necessary)
- Carbon/Graphite (labels can be attached to the metric name with
dot-separation)
- StatsD (see graphite when assuming that graphite is used as
backend, which is the default)
The util.statsd module has been ported to use the OpenMetrics
model as a proof of concept. An implementation which exposes
the util.statistics backend data as Prometheus metrics is
ready for publishing in prosody-modules (most likely as
mod_openmetrics_prometheus to avoid breaking existing 0.11
deployments).
At the same time, the previous measure()-based API had one major
advantage: It is really simple and easy to use without requiring
lots of knowledge about OpenMetrics or similar concepts. For that
reason as well as compatibility with existing code, it is preserved
and may even be extended in the future.
However, code relying on the `stats-updated` event as well as
`get_stats` from `statsmanager` will break because the data
model has changed completely; in case of `stats-updated`, the
code will simply not run (as the event was renamed in order
to avoid conflicts); the `get_stats` function has been removed
completely (so it will cause a traceback when it is attempted
to be used).
Note that the measure_*_event methods have been removed from
the module API. I was unable to find any uses or documentation
and thus deemed they should not be ported. Re-implementation is
possible when necessary.
[0]: https://openmetrics.io/
[1]: https://prometheus.io/
[2]: #959
[3]: #960
When set, no periodic statistics collection is done by
core.statsmanager, instead some module is expected to call collect()
when it suits. Obviously only one such module should be enabled.
Quoth jonas’
> correct way is to scrape the internal sources on each call to /metrics
> in the context of Prometheus
"manual" as opposed to "automatic", from the point of view of
statsmanager.
Zash> Btw, this conditional and loop, shouldn't it be covered by the timing measurement?
Zash> Isn't that where all the util.statistics work is done?
MattJ> Yeah, it should
Zash> ("the", but there's two ... which one‽)
MattJ> Yeah... not sure :)
MattJ> Processing I guess
